Video Captioning on CaReBench Object
54.8RecallCapRL++
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| CapRL++Backbone=Qwen3-VL-4B2026.06 | 54.8 | 70.6 | 61.7 | |
| Qwen3-VL-235B-A22BVariant=Instruct2026.06 | 32.4 | 80.3 | 46.1 | |
| Qwen3-VL-8BVariant=Instruct2026.06 | 32.2 | 79.2 | 45.8 | |
| Qwen3-VL-4BType=baseline2026.06 | 29.9 | 80.3 | 43.6 |